Systems and methods for distributing intent models

• US 10,452,782 B1
• Filed: 02/20/2018
• Issued: 10/22/2019
• Est. Priority Date: 02/20/2018
• Status: Active Grant

• Current Assignee: Meta Platforms, Inc. (f/k/a Facebook, Inc.)
• Original Assignee: Meta Platforms, Inc. (f/k/a Facebook, Inc.)
• Inventors: Kumar, Anuj, Dumoulin, Benoit F.
• Primary Examiner(s): Zhang, Leshui
• Application Number:
  US15/900,703
• Time in Patent Office: 609 Days
• Field of Search: 704 2, 704 7, 704 8, 704 9, 704E15014, 704250, 704255, 7042561, 7042564, 7042565, 7042566, 704259, 704266, 704263, 704E15046, 704E15047, 706 12, 706 13, 709217, 709218, 709219, 709229, 709244, 705 2635
• CPC Class Codes: G06F 40/295 Named entity recognition
• CPC Class Codes: G06F 40/30 Semantic analysis
• CPC Class Codes: G06N 20/00 Machine learning
• CPC Class Codes: G06Q 30/0601 Electronic shopping [e-shopping]

Abstract

Systems, methods, and non-transitory computer-readable media can receive, from a first entity, training data for training an intent model associated with a first intent of a plurality of intents. A first intent model associated with the first intent is generated based on the training data. The first intent model is made available in an intent marketplace for access by a second entity.

Claims

1. A computer-implemented method comprising:

receiving, by a computing system, from a first entity, training data for training an intent model associated with a first intent of a plurality of intents, wherein the training data comprises a plurality of natural language statements associated with the first intent;

generating, by the computing system, a first intent model associated with the first intent based on the training data, wherein the first intent model is trained to process natural language requests associated with the first intent;

making, by the computing system, the first intent model available in an intent marketplace for access by other entities to process natural language requests using the first intent model;

receiving, by the computing system, an indication that a second entity has purchased access to the first intent model in order to process natural language requests associated with the first intent and directed to the second entity; and

processing, by the computing system, a natural language request directed to the second entity using the first intent model based on a determination that the second entity has purchased access to the first intent model.

2. The computer-implemented method of claim 1, wherein the training data comprises a set of private training data and a set of public training data.

3. The computer-implemented method of claim 2, wherein the first intent model is a public intent model generated based on the set of public training data without utilizing the set of private training data.

4. The computer-implemented method of claim 3, further comprising generating a second intent model associated with the first intent based at least on the set of private training data.

5. The computer-implemented method of claim 4, wherein the first intent model is made available in the intent marketplace, and the second intent model is not made available in the intent marketplace.

6. The computer-implemented method of claim 1, further comprising:

testing the first intent model based on a set of testing data; and

publishing results of the testing the first intent model in the intent marketplace.

7. The computer-implemented method of claim 6, wherein

the set of testing data comprises a plurality of natural language user requests, and

the testing the first intent model comprises processing the plurality of natural language user requests based on the first intent model.

8. A system comprising:

at least one processor; and

a memory storing instructions that, when executed by the at least one processor, cause the system to perform a method comprising;

receiving, from a first entity, training data for training an intent model associated with a first intent of a plurality of intents, wherein the training data comprises a plurality of natural language statements associated with the first intent;

generating a first intent model associated with the first intent based on the training data wherein the first intent model is trained to process natural language requests associated with the first intent;

making the first intent model available in an intent marketplace for access by other entities to process natural language requests using the first intent model;

receiving an indication that a second entity has purchased access to the first intent model in order to process natural language requests associated with the first intent and directed to the second entity; and

processing a natural language request directed to the second entity using the first intent model based on a determination that the second entity has purchased access to the first intent model.

9. The system of claim 8, wherein the training data comprises a set of private training data, and a set of public training data.

10. A non-transitory computer-readable storage medium including instructions that, when executed by at least one processor of a computing system, cause the computing system to perform a method comprising:

receiving, from a first entity, training data for training an intent model associated with a first intent of a plurality of intents, wherein the training data comprises a plurality of natural language statements associated with the first intent;

generating a first intent model associated with the first intent based on the training data wherein the first intent model is trained to process natural language requests associated with the first intent;

making the first intent model available in an intent marketplace for access by other entities to process natural language requests using the first intent model;

receiving an indication that a second entity has purchased access to the first intent model in order to process natural language requests associated with the first intent and directed to the second entity; and

processing a natural language request directed to the second entity using the first intent model based on a determination that the second entity has purchased access to the first intent model.

11. The non-transitory computer-readable storage medium of claim 10, wherein the training data comprises a set of private training data, and a set of public training data.